====== Add 2 numbers =====
1. At the command prompt (in SCC) type (tocreate ptx file for matlab):
	nvcc -ptx add.cu

2. To specify the entry point for MATLAB kernel, run(at matlab prompt):
	k = parallel.gpu.CUDAKernel('add.ptx', 'addnums.cu');

3. Run kernel (kernel takes 2 arguments):
	out = feval(k, 7, 21);

====== Add 2 vectors =====
1. At the command prompt (in SCC) type (tocreate ptx file for matlab):
	nvcc -ptx add.cu

2. To specify the entry point for MATLAB kernel, run(at matlab prompt):
	k = parallel.gpu.CUDAKernel('add.ptx', 'addvecs.cu');

3. Run kernel:
	N = 128;
	k.ThreadBlockSize = N;
	feval(k, ones(N, 1), ones(N, 1));